SPE 84226

A Mechanistic Heat Transfer Model for Vertical Two-Phase Flow

Ryo Manabe, SPE, Japan National Oil Corp., Qian Wang, SPE, Hong-Quan Zhang, SPE, Cem Sarica, SPE and James P.
Brill, SPE, The University of Tulsa
Copyright 2003, Society of Petroleum Engineers Inc.
This paper was prepared for presentation at the SPE Annual Technical Conference and
Exhibition held in Denver, Colorado, U.S.A., 5 8 October 2003.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300
words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract
Convective heat transfer for vertical gas-liquid two-phase flow
was investigated experimentally and theoretically.
Experimental data on convective two-phase heat transfer
were acquired with a crude oil-natural gas system at cooling
conditions using a large diameter (2.067-in I.D.), high
pressure (450 psia) test facility. Flow pattern dependencies of
convective heat transfer with changing liquid and gas
velocities were revealed.
A comprehensive mechanistic heat transfer model was
developed by flow pattern dependent approach for bubbly,
intermittent and annular flow in vertical pipes. The model is
capable of predicting flow pattern first and then predicting
hydrodynamics and heat transfer based on the
predicted flow-pattern.
Comparing with experimental data, the model is found to
predict two-phase flow heat transfer coefficient within 26%
error for all flow patterns showing a better overall
performance than existing correlations.
Introduction
As oil and gas production moves to deepwater environments,
production systems with subsea completions and tiebacks to
existing platforms have become common occurrences. In such
systems, long-distance transportation of unprocessed reservoir
fluids, which are normally multiphase systems, from the
reservoir to downstream process facilities, must be assured. In
order to optimize the design and operation of such systems,
engineers must understand how petroleum fluids behave, both
hydrodynamically and thermally, during transportation. In
particular, wax deposition during transportation of waxy
crudes is found to be sensitive to convective heat transfer1.
Experimental data2-5 on two-phase heat transfer were
reported for a variety of fluids, gas and liquid flow rates, pipe
diameter, inclination, and flow patterns. There are a number of

single-phase heat transfer correlations6 and prediction

methods2,3,6-9 available in the literature for two-phase heat
transfer. Convective heat transfer in gas-liquid two-phase flow
clearly depends on the resulting flow patterns under given
operating conditions. Therefore, the prediction models must be
able to predict the resulting flow pattern first, and predict the
hydrodynamics and heat transfer for the specific flow pattern.
The Tulsa University Paraffin Deposition Project
(TUPDP) developed a prediction method for two-phase heat
transfer that combined two-phase flow pattern and
hydrodynamics mechanistic models and flow-patterndependent heat transfer correlations1. Based on a comparison
with a set of published experimental data, Aggour3 correlation
for bubbly flow; Rezkallah-Sims5 correlation for intermittent
flow and Ravipudi-Godbold7 correlation for annular flow were
recommended by Kim et al.10.
Multiphase flow modeling moved from correlation
approach to mechanistic approach. Although comprehensive
flow pattern and hydrodynamics models are available11-13, no
adequate heat transfer models exist. Moreover, no available
experimental data on two-phase heat transfer in crude oil natural gas systems with cooling conditions under high
pressure conditions were found. Mechanistic models for
improved prediction of two-phase convective heat transfer
are needed.
In this paper, a comprehensive mechanistic model for
vertical upward two-phase flow convective heat transfer is
proposed. The following sections present a summary of the
experimental and modeling studies.
Experimental Study
Test Facility and Heat Transfer Measurement Method
South Pelto crude oil (35o API gravity) was used as the liquid
phase. Natural gas supplied by Oklahoma Natural Gas
Company was used as the gas phase. The Prandtl numbers of
the liquid and gas phases were 32 < PrL < 42 , and Prg 0.74 ,
respectively.
A high-pressure multiphase test facility shown in Fig. 1
was utilized for the experimental study. A schematic diagram
of the test facility is shown in Fig. 2. Figure 3 shows the test
section for heat transfer measurements. A hot crude oil-natural
gas mixture (about 140oF) was flowed in the internal pipe of
the jacketed test section, and chilled glycol mixture (120oF)
was flowed counter-currently in the jacket to maintain a
constant cold pipe wall temperature. Therefore, heat loss for
the crude oil-natural gas mixture occured over the entire test

SPE 84226

section. The inlet and outlet temperatures for both the twophase mixture and glycol were measured during the test
duration after the flow became steady hydrodynamically
and thermally.
Average convective heat transfer coefficients are
determined by a heat balance based on heat flux
measurements14. The total heat loss of the two-phase mixture
between the inlet and the outlet of the measurement interval,
Q, is determined by
Q = (m& L C pL + m& g C pg )Tm . ........................................ (1)

Where, Tm represents the difference in bulk temperatures of

the twophase mixture between inlet and outlet. Once the total
heat loss, Q, is obtained, the overall heat transfer coefficient of
the system, U, is calculated by:
U=

(dL )TLM

. .......................................................... (2)

Where, TLM represents the logarithmic mean temperature

difference between the two-phase mixture and glycol. The
average convective two-phase heat transfer coefficient, hTP,
can then be obtained from:
r
r
1
1
1 ri
....................................... (3)
=
+ i ln o +
U hTP k p
ri h gl ro

The system pressure was fixed at about 350 psig. The inlet
glycol temperature was fixed at 120oF. The inlet temperature
of the two-phase mixture was set as high as possible to reduce
measurement uncertainty. The fluid inside temperatures varied
from 139oF to 157oF. The calculated relative uncertainties in
the convective two-phase heat transfer coefficient varied
between 2.5 and 13.0 percent of the reading14.
Experimental Results
A total of 55 tests were carried out, including 5 tests for
single-phase liquid, 6 tests for single-phase gas, 8 tests for
bubbly flow, 27 tests for intermittent flow, and 9 tests for
annular flow.
Single-Phase Liquid Flow. The experimental convective
heat transfer coefficients for single-phase liquid flow, hSL,
were measured to verify the facility and evaluate the existing
model. The convective heat transfer coefficients increase
linearly with increasing superficial liquid velocity, except for
the data point in laminar flow (ReL=828).
Single-Phase Gas Flow. The experimental convective heat
transfer coefficients for single-phase gas flow, hSg, were
measured. All the data were acquired under a fully-developed
turbulent flow (Reg > 40664). The convective heat transfer
coefficients increase linearly with increasing superficial
gas velocity.
Vertical Two-Phase Flow. Experimental convective twophase heat transfer coefficients, hTP, are plotted against
superficial gas velocities at fixed superficial liquid velocities
in Fig. 4. In general, higher hTP values are obtained for higher
vSL values for a given vSg.

In the bubbly flow region, hTP increases slightly with

increasing vSg at fixed vSL. In the intermittent flow region,
however, hTP remains almost constant as vSg increases until the
intermittent-annular transition boundary is reached. At the
intermittent-annular transition boundary, hTP increases at
relatively low vSL and decreases for relatively high vSL. In the
annular flow region, hTP increases again as vSg increases for a
given vSL.
The enhancements of convective heat transfer coefficient
in two-phase flow due to the addition of a gas phase, defined
as hTP/hSL, are plotted against the gas-liquid superficial
velocity ratio, vSg/vSL in Fig. 5. hTP/hSL values increase slightly
in bubbly and intermittent flow, and increase significantly in
annular flow.
Modeling Study
A comprehensive two-phase heat transfer model was
developed based on a mechanistic approach. The model
consists of a flow pattern prediction model and a set of
individual mechanistic models for predicting hydrodynamics
and heat transfer. An existing mechanistic model, Ansari et
al.11, for vertical two-phase flow, is used to predict the flow
pattern and hydrodynamics. The flow pattern dependent
mechanistic heat transfer models are newly developed.
Single-Phase Heat Transfer
Laminar Flow. For high Prandtl number fluids, such as
crude oil, hydrodynamically developed but thermally
developing flow is a reasonable assumption. The Sieder-Tate
correlation15 for a constant wall temperature boundary
condition is widely used.
d

Nu = 1.86 Re Pr
L

1/ 3

0.14

............................... (4)

Turbulent Liquid Flow. When the difference between the

fluid bulk and the wall temperatures is significant, the effect of
the physical property variations on heat transfer must be
considered. Sieder-Tate15 and Petukhov-Kirillov16 proposed
the
following
correlations,
respectively,
for
cooling conditions,
Nu = 0.023 Re

45

Pr b
w
1 3

f
Re Pr
8
Nu =
f
1.07 + 12.7
Pr 2 / 3 1
8

0.14

.................................. (5)

0.25

........................... (6)

Turbulent Gas Flow. The most widely used correlation for

fully-developed turbulent flow with constant properties under
cooling conditions was recommended by Dittus-Boulter17 as,
Nu g = 0.023 Re g 0.8 Pr g 0.3 ............................................ (7)

SPE 84226

Bubbly Flow
A pseudo single-phase approach is used to predict the
convective heat transfer coefficient, as well as hydrodynamics,
for bubbly flow. The flow is assumed to be fully turbulent and
the heat transfer in the gas phase is neglected. Moreover, the
bubbles are assumed to have no influence on the boundary
layer and the increase in hTP is only due to the increase in the
mean velocity of the two-phase mixture.
The Petukhov-Kirillov correlation is utilized with liquid
physical properties, L, L, CpL and kL, and mixture
velocity, vm,
NuTP

f
ReTP PrL
8
=
f
1.07 + 12.7
PrL 2 / 3 1
8

L ,b

L, w

0.25

. ................. (8)

Where the two-phase Reynolds number, ReTP, is defined as,

Re TP =

L vm d
............................................................. (9)
L

The convective heat transfer coefficient is obtained as,

hTP =

Nu TP k L
............................................................ (10)
d

Annular Flow
Kaminsky9 proposed the modeling of convective heat transfer
separately for turbulent and laminar. In this study, Kaminskys
method is utilized only for annular flow liquid films with
slight modifications. The definition of the critical liquid film
Reynolds number, ReF, for the laminar-turbulent transition is
still an unsolved issue. Many studies have been carried out on
this topic. It was found that the critical liquid film Reynolds
number varied in the range between 1,000 and 1,200.18 In this
study, a value of the critical liquid film Reynolds number of
ReF = 1100 is used.
Turbulent Liquid Film. The Prandtl-Taylor analogy6 is
applied for the turbulent liquid film with the
following assumptions:
Heat transfer between the liquid film and gas core
is negligible.
Flow in the film is divided into two parts: laminar
boundary layer and turbulent core.
Laminar boundary layer is much thinner than the
liquid film thickness and there are no bubbles in the
liquid film.
Temperature drop occurs only in the laminar
boundary layer.
q/ is constant over laminar boundary layer.
The momentum transfer and heat transfer in laminar
boundary layer are expressed as, respectively;

v
dv
= lb , ........................................................ (11)
lb
dy

q = k

T Tw
dT
............................................ (12)
= k lb
dy
lb

Taking the ratio, q/, and integrating over a laminar boundary

layer, we obtain,
q

qw

k
= L
L

(Tlb Tw )

....................................... (13)
vlb

or
hTP =

k
qw
= L
Tw Tb L

vlb

. .................................... (14)

Here, the fluid velocity at the edge of the laminar boundary

layer, vlb, is estimated as,
1/ 2

vlb = w . ........................................................... (15)

Where is a constant depending on the exact definition of

edge of laminar boundary layer, and is generally taken as
equal to 519.
Sustituting Eq. (15) into Eq. (14), the convective heat
transfer coefficient, hTP, can be expressed as,
1 kL
hTP =
L F ................................................... (16)
5 L
The film shear stress can be estimated as,

F =

d dP
........................................................ (17)

4 dL TP , f

Substituting Eq. (17) into Eq. (16), we obtain;

hTP

k L L dP
=
d

10 L dL TP , f

1/ 2

.................................. (18)

The convective heat transfer coefficient for superficial

liquid film, hSF, can be expressed in terms of superficial
pressure gradient in the same manner,

hSF

k L L dP
=
d

10 L dL SL, f

1/ 2

.................................. (19)

Taking the ratio of the two-phase and superficial liquid

film coefficients gives

SPE 84226

hTP
h SF

dP

dL TP. fric

=
dP

dL SL, fric

1/ 2

= L ..................................... (20)

Where L is the liquid phase Lockhart and Martinelli

parameter20. Finally, the following expression for the
convective heat transfer coefficient in turbulent liquid film
annular flow is obtained as,
hTP = hSF L ................................................................ (21)

Here, hSF can be estimated by the superficial liquid convective

heat transfer correlation, namely the Petukhov-Kirillov
correlation, with liquid fluid properties and the superficial
liquid film velocity, vSF.
Laminar Liquid Film. No bubbles in the liquid film are
assumed. Heat transfer between the liquid film and gas core is
neglected. Moreover, heat transfer is assumed to be equivalent
to that of single phase laminar flow at the same average
velocity in a pipe with an equivalent hydraulic diameter.
The convective heat transfer coefficient for laminar liquid
film annular flow can be expressed, based on the Sieder-Tate
correlation, as
hTP =

= 1.86 ann

hTP = hTB + h LS (1 ) . ............................................ (25)

Taylor Bubble Region. Taylor bubble region can be

considered as a falling liquid film. Effective single-phase
approach, as in the laminar annular flow model, is used to
model heat transfer in the Taylor bubble region.
The convective heat transfer coefficient in the Taylor
bubble region, hTB, can be expressed as,
hTB = 2.08 H L 2 / 3 h LTB . .............................................. (26)

Liquid Slug Body. A pseudo single-phase approach is used

to model the heat transfer in the liquid slug body, as same as
the bubbly flow model.
The Nusselt number in the liquid slug body, NuLS, can be
expressed as,

Nu LS

Nu TP k L
d hF
kL
d
Re SF PrL
d
L

Intermittent Flow
The convective heat transfer in vertical intermittent flow can
be analyzed separately for the heat transfer coefficient in the
liquid slug body, hLS, and in the Taylor bubble region, hTB.
Thus, the average coefficient over a slug unit, hTP, is
expressed as,

1/ 3

. ... (22)

0.14

H LF

2 / 3

Where HLF represents liquid film holdup and ann is a shape

factor for annular liquid film heat transfer.
Taking the ratio of the two-phase and superficial liquid
film heat transfer coefficients, the following equation
is obtained,
hTP ann
=
H LF 2 / 3 . ................................................. (23)
cir
h SF

Here, the ratio of shape factors, ann/circ, can be considered as

the ratio of Nusselt numbers of a thin liquid film and a circular
pipe for fully developed laminar flow. The Nusselt number for
fully developed laminar thin liquid film is 7.60 while for fully
developed laminar circular pipe flow it is 3.66. Therefore, the
expression for convective heat transfer coefficient in laminar
liquid film annular flow is,
hTP 7.60
=
H LF 2 / 3 = 2.08 H LF 2 / 3 . ........................ (24)
h SF 3.66

The convective heat transfer coefficient for superficial

liquid film is calculated by the Sieder-Tate correlation with
liquid physical properties and liquid film Reynolds
number, ReSF.

f
Re LS PrLS
8
=
f
1.07 + 12.7
PrLS 2 / 3 1
8

LS ,b

,
LS
w

0.25

. ......... (27)

Where the Reynolds number for the liquid slug body, ReLS,
and the Prandtl number of the liquid slug body, PrLS, are
respectively defined as,
Re LS =

PrLS =

LS v m d
,........................................................ (28)
LS

C PLS LS
. ....................................................... (29)
k LS

Because the laminar boundary layer is completely

disturbed by mixing in the front of the liquid slug body, the
twophase physical properties, TP, TP, CpTP and kTP, are used
instead of the liquid properties.
The convective heat transfer coefficient in the liquid slug
body is expressed as,
h LS =

Nu LS d
. ........................................................... (30)
k LS

Model Evaluation
The performance of the mechanistic two-phase heat transfer
model is evaluated using both sensitivity and statistical error
analyses. The model is also compared with the existing
correlations recommended by Kim et al.10(OSU Model). The
statistical error analysis results are summarized in Tables 1-4
and discussed separately for respective flow patterns.

SPE 84226

Six different statistical parameters 1-6 are reported in

Tables 1-4. The average relative error, 1, and the average
error, 4, are measures of the agreement between the measured
data and predicted results. They indicate the degree of over
prediction (positive values) or under prediction (negative
values). The degree of agreement between the measured and
predicted data is reflected by the average absolute relative
error, 2, and the average absolute error, 5, because the
negative and positive errors do not cancel out in the
calculations. The standard deviations, 3 and 6, indicate the
scatter of the errors. The definition of these statistical
parameters can be found in Manabe14.
Single Phase Flow
The comparison between predicted and experimental
convective heat transfer coefficients for single phase liquid
flow is shown in Fig. 6. It is seen that the Sieder-Tate
correlation underpredicts the heat transfer coefficients and the
Petukhov-Kirillov correlation performs better.
Comparison between the predicted and experimental
convective heat transfer coefficients for single-phase gas flow
is shown in Fig. 7. Good agreement is obtained over the entire
range of the data. The results show that the Dittus-Boelter
correlation gives accurate predictions of the convective heat
transfer coefficients for single-phase gas flow in this system.
The results also confirm the reliability of the experimental
data acquired from the test facility of this study.
Bubbly Flow
Both predicted and experimental convective heat transfer
coefficients are plotted against superficial gas velocities at
different liquid superficial velocities in Fig. 8. Both predicted
and experimental values of hTP increase as vSg increases,
except for the data point near the bubbly-intermittent
transition boundary (vSL=4.97 ft/s, vSg=1.09 ft/s).
The predicted and experimental enhancements of
convective heat transfer coefficients, hTP/hSL, are plotted
against the ratio of superficial gas and liquid velocities,
vSg/vSL, at different vSL, in Fig. 9. Both predicted and
experimental hTP/hSL increase as vSg/vSL increases. It was
found that hTP/hSL is essentially independent of vSL.
Comparisons between predicted and experimental hTP
values are shown in Fig. 10 and Table 1. hTP values are
predicted quite well by both the newly developed mechanistic
model and the Aggour3 correlation. Both models predict hTP
within 20% uncertainity. The mechanistic model tends to
under predict the convective heat transfer coefficients, as in
the case of single-phase liquid flow.
Annular Flow
Predicted and experimental hTP values are plotted against vSg
at different values of vSL in Fig. 11. Predicted values of hTP
decrease as vSg increases at some vSL, even though the
experimental hTP values increase as vSg increases. The reason
for the inconsistency in sensitivity of hTP to vSg between the
predicted and experimental values is as follows. As vSg
increases, the liquid entrainment fraction in the gas core
increases. Thus, the superficial liquid film velocity, vSF,
decreases and the calculated hSF decreases.

Predicted and experimental values of hTP/hSF are plotted

against vSg/vSL, at different vSL in Fig. 12. Experimental
hTP/hSF values increase as vSg increases for all liquid
superficial velocity. However, at some liquid superficial
velocity, the predicted hTP/hSL decreases as vSg increases.
Comparisons between the predicted and experimental
values of hTP are shown in Fig. 13. The mechanistic model can
predict hTP within 30% uncertainity for both turbulent and
laminar liquid film.
Intermittent Flow
As shown in Figs. 14 and 15, predicted and experimental
values of hTP and hTP/hSL, are plotted against vSg and vSg/vSL at
different vSL, respectively. At relatively low vSL, namely the
cases of vSL=0.20, both the predicted and experimental hTP and
hTP/hSL values slightly increase as vSg and vSg/vSL increase. At
medium vSL, namely the cases of vSL=0.5, and vSL=1.0 ft/s,
both the predicted hTP and hTP/hSL values are fluctuated as vsg
and vSg/vSL increase. At relatively high vSL, namely the case
with vSL=5.00 ft/s, both the predicted and experimental values
of hTP and hTP/hSL increase as vSg and vSg/vSL increase.
Comparisons between predicted and experimental values
of hTP are shown in Fig. 16 and Table 3. hTP is predicted by
both the newly developed mechanistic model and the
Rezkallah-Sims8 correlation. The mechanistic model underpredicts the convective heat transfer coefficients, in particular
at high vSL, within 30% uncertainity. The Rezkallah-Sims
correlation fails to predict at high vSL.
Overall Performance
Comparison between predicted and experimental values of
hTP for all the vertical flows is shown in Fig. 17. The results of
the statistical evaluation are shown in Table 4. The newly
developed mechanistic model can predict hTP within 26% error
for vertical flows.
Concluding Remarks
In this study, convective heat transfer for gas-liquid two-phase
flow has been investigated experimentally and theoretically.
Based on the experimental study, it was found that hTP
increased slightly for bubbly flow and significantly for annular
flow, respectively, with increase of vSg for a constant vSL. For
intermittent flow, hTP remained constant with increasing vSg at
a given vSL. For high vSL values, hTP decreased near the
intermittent-annular transition boundary.
A comprehensive mechanistic heat transfer model was
developed by flow pattern dependent approach for bubbly,
intermittent and annular flows in vertical pipes. The pseudo
single-phase approach is used for bubbly flow. Prandtl-Taylor
analogy in combination with the Lockhart-Martinelli
parameter, for annular flow with turbulent liquid film is
applied while Ravipudi-Godbold correlation for annular flow
with laminar liquid film is utilized. Weighted average method,
the pseudo single-phase approach for liquid slug body and the
effective single-phase liquid flow approach for liquid film
region are employed for intermittent flow.
The proposed mechanistic model performed better than
the existing correlations. The model can predict hTP within 20,
30 and 25 percent uncertainity for bubbly, annular and
intermittent flows, respectively.

SPE 84226

Nomenclature
Cp = heat capacity
d = pipe diameters
f = friction factor
HL = Liquid Holdup
h = convective heat transfer coefficient
k = thermal conductivity
L = length
m = mixture
m& = mass flow rate
Nu = Nusselt number
Pr = Prandtl number
q = heat flux
Q = heat flow
r = pipe radius
Re = Reynolds number
U = overall heat transfer coefficient
y = distance
= ratio of Taylor bubble to slug unit length
= shape factor (constant)
= dimensionless pressure group defied by Lockhart
and Martinelli
= viscosity
= density
= shear stress
Subscript
ann. = annular
b = bulk
circ. = circular
F = film
g = gas
gl = glycol
i = inside
L = liquid
lb =at laminar boundary layer
LS = liquid slug
LTB = liquid in Taylor bubble region
o = outside
p = pipe
SF = superficial film
Sg = superficial gas
SL = superficial liquid
TB = Taylor bubble
TP = two-phase
w = wall
Acknowledgments
Authors thank Tulsa University Fluid Flow Project (TUFFP)
and Tulsa University Paraffin Deposition Project (TUPDP)
member companies for their supports. They also acknowledge
Tony Butler and Mr. Emmanuel Delle Case for their
dedicated work.
References
1.
2.

Matzain, A.: Multiphase Flow Paraffin Deposition

Modeling,Ph.D. Dissertation, U. of Tulsa (1999).
Knott, R.F., Anderson, R.N., Acrivos, A. and Petersen, E.E.:
An Experimental Study of Heat Transfer to Nitrogen-Oil
Mixtures, Ind. and Eng. Chem., 51, 1369-1372. (1959).

3.
4.
5.
6.
7.
8.

9.
10.

11.
12.
13.
14.
15.
16.

17.

Aggour, M.A.: Hydrodynamics and Heat Transfer in TwoPhase Two-Component Flow, Ph.D. Dissertation, U. of
Manitoba, Canada. (1978).
Vijay, M.M.: A Study of Heat Transfer in Two-Phase TwoComponent Flow in a Vertical Tube, Ph.D. Dissertation, U. of
Manitoba (1978).
Rezkallah, K.S.: Heat Transfer and Hydrodynamics in TwoPhase Two-Component Flow in a Vertical Tube, Ph.D.
Dissertation, U. of Manitoba (1987).
Kakac, S. and Yener, Y.: Convective Heat Transfer, 2nd Edition,
CRC Press (1995).
Ravipudi, S.R. and Godbold, T.M.: The Effect of Mass
Transfer on Heat Transfer Rates for Two-Phase Flow in Vertical
Pipe, Proc. 6th Int. Heat Transfer Conf., 1, 505-510 (1978).
Rezkallah, K.S. and Sims, G.E.: An Examination of
Correlations of Mean Heat Transfer Coefficients in Two-Phase
and Two-Component Flow in Vertical Tubes, AIChE Symp.
Series, 83, 109-114. (1987).
Kaminsky, R.D.: "Estimation of Two-Phase Flow Heat Transfer
in Pipes," ASME J. Energy Res. Tech., 121, 75-80 (1999).
Kim, D., Ghajar, A.J., and Dougherty, R.L.: Development of
Improved Two-Phase Two-Component Pipe Flow Heat Transfer
Correlations from Existing Correlations and Published Data,
presented at ASME/JSME Joint Thermal Engineering
Conference (1999).
Ansari, A.M., Sylvester, N.D., Sarica, C., Shoham, O., and Brill,
J.P.: A Comprehensive Mechanistic Model for Upward Flow in
Pipes, SPE J. Prod. & Fac., 143-152 (May 1994).
Xiao, J.J., Shoham, O. and Brill, J.P.: A Comprehensive
Mechanistic Model for Two-Phase Flow in Pipelines, SPE
20631, presented at 1990 SPE ATCE, New Orleans (1990).
Kaya, A.S., Sarica, C. and Brill, J.P.: Comprehensive
Mechanistic Modeling of Two-Phase Flow in Deviated Wells,
SPE 56522, presented at 1999 SPE ATCE, Houston (1999).
Manabe, R.: A Comprehensive Mechanistic heat Transfer
Model for Two-Phase Flow with High-Pressure Flow Pattern
Validation, Ph.D. Dissertation, U. of Tulsa (2001).
Sieder, E.N. and Tate, G.E.: Heat Transfer and Pressure Drop
of Liquids in Tubes, Ind. Eng. Chem, 28, 1429-1453. (1936).
Petukhov, B.S.: Heat Transfer and Friction in Turbulent Pipe
Flow with Variable Physical Properties, Advances in Heat
Transfer, 6, 505-564 (1970).
Dittus, F.W. and Boelter, L.M.K.: Heat Transfer in Automobile
Radiators of the Tubular Type, Univ. Calif. Pub. Eng., 2,
443 (1930).

18. Hewitt, G.F.: Annular Two Phase Flow, Pergamon

Press (1970).
19. Schlichting, H.: Boundary-Layer Theory, 7th EditionEnglish, McGraw-Hill Books Inc. (1979).
20. Shoham, O.: Two-Phase Flow Modeling, U. of
Tulsa (1999).
SI Metric Conversion Factors
bbl x 1.589 873
E-01 = m3
*
cp x 1.0
E-03 = Pa.s
ft x 3.048*
E-01 = m
ft2 x 9.290 304* E-02 = m2
ft3 x 2.831 685
E-02 = m3
lbm x 4.535 924
E-01 = Kg
psi x 6.894 757 E+00 = kPa
Conversion factor is exact

SPE 84226

Table 1. Statistical Evaluation of Vertical Bubbly Heat Transfer Model

1,hTP

2,hTP

3,hTP

4,hTP

5,hTP

6,hTP

(-)

(-)

(-)

(Btu/hr-sq.ft-F)

(Btu/hr-sq.ft-F)

(Btu/hr-sq.ft-F)

This model

-0.157

0.157

0.038

-27.041

27.041

14.709

OSU Model

-0.139

0.139

0.034

-23.178

-23.178

11.202

Table 2. Statistical Evaluation of Vertical Annular Heat Transfer Model

This model
OSU Model

1,hTP

2,hTP

3,hTP

4,hTP

5,hTP

6,hTP

(-)

(-)

(-)

(Btu/hr-sq.ft-F)

(Btu/hr-sq.ft-F)

(Btu/hr-sq.ft-F)

-0.147
-0.092

0.3
0.188

0.312
0.198

-19.27
-17.222

48.083
33.416

50.555
67.628

Table 3. Statistical Evaluation of Vertical Intermittent Heat Transfer Model

1,hTP

2,hTP

3,hTP

4,hTP

5,hTP

6,hTP

(-)

(-)

(-)

(Btu/hr-sq.ft-F)

(Btu/hr-sq.ft-F)

(Btu/hr-sq.ft-F)

This model

-0.203

0.221

0.177

-20.439

22.053

19.867

OSU Model

0.336

0.434

0.637

54.471

64.555

135.094

Table 4. Statistical Evaluation of Vertical Flows Heat Transfer Model

1,hTP

2,hTP

3,hTP

4,hTP

5,hTP

6,hTP

(-)

(-)

(-)

(Btu/hr-sq.ft-F)

(Btu/hr-sq.ft-F)

(Btu/hr-sq.ft-F)

Bubbly

-0.157

0.157

0.038

-27.041

27.041

14.709

Intermittent
Annular
Overall

-0.203
-0.147
-0.164

0.221
0.3
0.258

0.177
0.312
0.26

-20.439
-19.27
-17.139

22.053
48.083
34.375

19.867
50.555
40.999

Figure 1. Test Facility

Figure 2. Process Flow Diagram of Test Facility

SPE 84226

TT4
5

TT 50

TT40

TT30

TT35

250

Glycol

1/2

200

TW1
TT49

TW 3

2 o

TT29
2.00 in.

TW2
TF
1
45o

Predicted hSL (Btu/hr-ft - F)

1/2

1.38 in.

TT4
4

TF
2

TT39
0.69 in.

TW4

TF
3

150

100

TT 34

Laminar - Sieder-Tate
Turbulent - Petukhov-Kirillov
Turbulent - Sieder-Tate
0%
-10%
10%

50
0.20 in.

TW
5
0
0

50

100

150

200

250

2 o

0.197 ft
(0.06 m)

5.0197 ft
(1.53 m)

5.0197 ft
(1.53 m)

5.0197 ft
(1.53 m)

Experimental hSL (Btu/hr-ft - F)

5.0197 ft
(1.53 m)

Figure 6. Comparison between Predicted and Experimental

Convective Heat Transfer Coefficients for Single-Phase
Liquid Flow

Figure 3. Test section for heat transfer measurements

350.00

150

300.00

Annular

Intermittent

Bubbly

+10%

250.00

Predicted hTP (Btu/hr-ft - F)

120

-10%

hTP

2 o

200.00

150.00

VSL=O,2

100.00

VSL=0.5
VSL=1.0

90

60

VSL=2.0

50.00

VSL=2.5

30

VSL=4.0
VSL=5.0

0.00
0.10

1.00

10.00

100.00
0

v Sg (ft/sec)

Figure 4. Experimental Convective Heat Transfer Coefficients vs.

Superficial Gas Velocity for Vertical Two-Phase Flow

30

60

90

120

150

Experimental hTP (Btu/hr-ft2-oF)

Figure 7. Comparison between Predicted and Experimental

Convective Heat Transfer Coefficients for Single-Phase Gas Flow

100.00

300

hTP/hSL

hTP (Btu/hr-ft -F)

250

10.00

VSL=O,2

VSL=0.5

VSL=1.0

200
150
100

This model (vSL=2.0ft/ s)

Exp. data (vSL=2.0f t/ s)
this model (vSL=4.0ft /s)

VSL=2.0

VSL=2.5

VSL=4.0

50

Exp.data (vSL=4.0f t/ s)
This model (vSL=5.0f t/s)
Exp. data (vSL=5.0ft /s)

VSL=5.0
1.00
0.01

0.10

1.00

10.00

100.00

1000.00

vSg /vSL

Figure 5. Enhancement of Convective Heat Transfer Coefficient

vs. Superficial Gas/Liquid Velocity Ratio for Vertical
Two-Phase Flow

0
0.10

1.00

10.00

v Sg (ft/sec)

Figure 8. Sensitivity of Convective Heat Transfer Coefficient to

Superficial Gas Velocity for Vertical Bubbly Flow

SPE 84226

100

10

This model ( vSL=0.2f t / s)

Exp. dat a ( vSL=0.2f t / s)
This model ( vSL=0.5f t / s)
Exp. dat a ( vSL=0.5f t / s)
t his model (vSL=1.0f t / s)
Exp. dat a ( vSL=1.0f t / s)

Exp.dat a ( vSL=2.5f t / s)

hTP/hSL,cal

hTP/hSL,cal

This model ( vSL=2.5f t / s)

This model ( vSL=5.0f t / s)

Exp.dat a ( vSL=5.0f t / s)

10

This model (vSL=2.0f t / s)

Exp. dat a (vSL=2.0f t / s)

This model (vSL=4.0f t / s)

Exp. dat a (vSL=4.0f t / s)

This model (vSL=5.0f t / s)

Exp. dat a (vSL=5.0f t / s)

0.1

0.01

0.1

v Sg/v SL
Figure 9. Sensitivity of Enhancement of Convective Heat Transfer
Coefficient to Gas/Liquid Superficial Velocity Ratio for Vertical
Bubbly Flow

10

vSg/vSL

100

1000

Figure 12. Sensitivity of Enhancement of Convective Heat

Transfer Coefficient for Gas/Liquid Superficial Velocity Ratio for
Vertical Annular Flow

400
300

200

Predicted hTP (Btu/hr-ft2-F)

Predicted h TP (Btu/hr-ft2-F)

250

This model
OSU Model
0%
+40%
-40%

350

This model
OSU model
0%
+30%
-30%

150
100

300
250
200
150
100
50

50

0
0

0
0

50

100
150
200
Experim ental hTP (Btu/hr-ft2-F)

250

100

150

200

250

300

Experimental hTP (Btu/hr-ft2-F)

300

Figure 10. Comparison between Predicted and Experimental

Convective Heat Transfer Coefficients for Vertical Bubbly Flow

50

Figure 13. Comparison between Predicted and Experimental

Convective Heat Transfer Coefficients for Vertical Annular Flow

350
400
This model (vSL=0.2f t /s)

hTP (Btu/hr-ft2-F)

300
250

Exp. data (vSL=0.2f t/ s)

This model (vSL=0.5f t/ s)
Exp.dat a (vSL=0.5f t /s)

250

This model (vSL=1.0ft / s)

h TP (Btu/hr-ft 2 -F)

350

300

Exp.dat a (vSL=1.0f t/ s)
This model (vSL=2.5f t/ s)
Exp. data (vSL=2.5ft / s)
This model (vSL=5.0f t/ s)

200

Exp.dat a(vSL=5.0ft / s)

150

150
100

100

50

50
0
1.00

200

Exp. data (vSL=0.2ft/s)

This model (vSL=0.2ft/s)
Exp.data(vSL=0.5ft/s)
This model (vSL=0.5ft/s)
Exp. data (vSL=1.0ft/s)
This model (vSL=1.0ft/s)
Exp.data (vSL=2.5ft/s)
This model (vSL=2.5ft/s)
Exp.data (vSL=5.0ft/s)
This model (vSL=5.0ft/s)

0
0.10
10.00

100.00

vSg (ft/sec)

Figure 11. Sensitivity of Convective Heat Transfer Coefficient to

Superficial Gas Velocity for Vertical Annular Flow

1.00

10.00

100.00

vSg (ft/sec)

Figure 14. Sensitivity of Convective Heat Transfer Coefficient to

Superficial Gas Velocity for Vertical Intermittent Flow

10

SPE 84226

hTP/hSL,cal

10

1
This model (vSL=0.2f t/ s)
Exp.data (vSL=0.2f t/ s)
This model (vSL=0.5f t/ s)
Exp.data (vSL=0.5f t/ s)
This model (vSL=1.0f t/ s)
Exp. dat a (vSL=1.0f t/ s)
This model (vSL=2.5f t/ s)
Exp. dat a (vSL=2.5ft/ s)
This model (vSL=5.0f t/ s)
Exp. dat a (vSL=5.0ft/ s)

0.1
0.1

10

100

v Sg/v SL
Figure 15. Sensitivity of Enhancement of Convective Heat
Transfer Coefficient to Gas/Liquid Superficial Velocity Ratio for
Vertical Intermittent Flow
1000
This model

Predicted hTP (Btu/hr-ft2-F)

900

OSU model

800

0%

700

+40%
-40%

600
500
400
300
200
100
0
0

100

200

300

400

Experimental hTP (Btu/hr-ft2-F)

Figure 16. Comparison between Predicted and Experimental

Convective
Heat
Transfer
Coefficients
for
Vertical
Intermittent Flow
400
Bubbly
Annular
Intermittent
0%
+40%
-40%

Predicted h TP (Btu/hr-ft -F)

350
300
250
200
150
100
50
0
0

50

100

150

200

250

300

350

400

Experimental h TP (Btu/hr-ft -F)

Figure 17. Comparison between Predicted and Experimental

Convective Heat Transfer Coefficients for Vertical Flows